When satellites have satellites:

The story of Iapetus

Kevin J. Walsh
Hal F. Levison
Amy C. Barr
Luke Dones (SwRI)
Steve Schwartz
Derek Richardson
(UMD)
Iapetus oddities
It appears that Iapetus has de-spun signicantly
in its lifetime,
Iapetus is currently rotating with a 79.3 day period,
synchronous with its orbital period.
Its global shape is not in hydrostatic equilibrium,
It has the shape of a body at 16 h rotation this is
the fossil bulge.
The equatorial ridge, ~15 km tall and ~50 km
wide, covers at least 110 of the equator.
Problems
de-spinning Iapetus
Saturnian tides fail when
operating on a rigid, cold
Iapetus,
De-spinning takes > 10 Gyrs
Saturnian tides can only
succeed with the aid of an
extra energy source,
Frequently modeled as
Short-/Long-lived radioactive
isotopes (i.e.
26
Al)
Equatorial ridge
Ridge formation by an
endogenic process is
challenging,
Stresses frequently are in the
wrong direction
Thermal bouyancy stress makes
it hard to lift such a large ridge
It appears to be one of oldest
features on the surface
The timing of its formation
must be considered.
What if Iapetus had a (sub-)satellite?
Tidal de-spinning
A satellite with n >
Iap
of
Iapetus evolves outward
An outwardly evolving satellite
(due to tides) de-spins Iapetus.
Larger sub-satellites take more
angular momentum from
Iapetus, de-spinning it faster
Smaller sub-satellites take
longer, affecting Iapetus less
Equatorial ridge
A small satellite is likely to
form from a debris disk
straddling the Roche limit,
Similar to Earth-Moon system
Debris inside the Roche limit/
Synchronous limit,
tidally evolves inward and,
gets pushed to the surface by
the satellite.
The in-falling ring could form
an equatorial ridge (Ip 2006).
Where could a satellite come from?
Collisional origin
Levison et al. (2010) propose
that a large impact into Iapetus
ejects signicant debris into
orbit
The disk quickly collisionally
damps to a disk.
Disk straddles the Roche
limit, producing a satellite
outside, leaving debris inside,
The collision would leave
Iapetus with a rotation rate
16h
Intact capture
Dombard et al. (2011)
propose the capture of a
satellite, following a
collision.
The proposed event would
require a ~100km satellite
(0.1% Iapetus mass),
Follow Canup (2005,2011)
to suggest orbits of:
a = 3.7-21 R
Iapetus

e = 0.1-0.8
Tidal evolution with a sub-satellite
As the sub-satellite evolves to
larger semi-major axis it will
eventually get stripped by
Saturn.
We found that sub-satellite
lifetimes drops precipitously
beyond a ~ 21 R
Iap
Orbital period at a ~ 21 R
Iap
is
~12 days.
For all following calculations
we consider this endpoint of
tidal evolution.
Initial semi-major axis (R
Iap
)
L
i
f
e
t
i
m
e

[
y
e
a
r
s
]

Tidal evolution with a sub-satellite
de-spinning and orbit
expansion rates depends on
the internal tidal dissipation
(Q/k
2
),
Q/k
2
is not static this ratio
depends on
The internal state
(Temperature, lithosphere
thickness )
Tidal frequency
We need to consider the
internal state of Iapetus as a
function of time.
Internal evolution of Iapetus
Castillo-Rogez (2007) and
Robuchon (2010) modeled the
internal evolution of Iapetus under
many effects
Initial porosity
Short-live radioactive isotopes
Long-lived radioactive isotopes
Convection (Robuchon only)
Castillo-Rogez (2007)
Robuchon (2010)
8
a
d
l
u
s

(
k
m
)

1lme (Myr)
A simple model of Iapetus
We needed a simple model
of Iapetus to calculate Q/k
2
at each tidal frequency.
200 km lithosphere with
high viscosity.
The rest of the interior is
either
Static =10
15
10
18
(t) according to Castillo-
Rogez (2007) or Robuchon
(2010)
We calculate Q/k
2
values for
this body as a function of the
internal
R
200 km 535 km
Tidal evolution for !=10
16
Pa s
Satellite o
rbital perio
d
Iapetus rotation period
q > 0.021
Sub-satellite slows Iapetus
rotation to < 12 days
(synchronous), before
reaching a>21 R
Iap
0.006 < q < 0.021
Sub-satellite evolves to
a>21 R
Iap
, and is stripped
by Saturn.

0.006 > q
Sub-satellite evolves too
slowly and never reaches
21 R
Iap .
It eventually
evolves back to Iapetus.

q=0.024 synchronous lock
q=0.018 sub-saLelllLe escape
q=0.004 sub-saLelllLe re-lmpacL
1lme (Myr)
l
a
p
e
L
u
s

s
p
l
n

a
n
d

s
u
b
-
s
a
L
e
l
l
l
L
e

o
r
b
l
L
a
l

p
e
r
l
o
d

(
d
a
y
s
)

De-spinning time-savings
Mass Ratio: q = M
ss
/M
Iap

T
i
m
e

[
M
y
r
]

For the simple cases with static internal viscosities sub-satellite mass
ratio q= M
ss
/M
iap
=0.021 de-spins 10x faster than by Saturn alone.


Open symbols: sub-satellite is too small is never stripped
Closed symbols: sub-satellite is stripped by Saturn
Mass Ratio: q = M
ss
/M
Iap

T
i
m
e

[
M
y
r
]

0.01 0.001 0.0001
De-spinning time-savings
For the Robuchon models with 72 ppb and 0.04 ppb
26
Al (SLRI)
q= M
ss
/M
iap
=0.021 de-spins > 10x faster than by Saturn alone.
For Castillo-Rogez (2007) model with LLRI and initial porosity
q=M
ss
/M
iap
=0.021 only de-spins 20% faster


Robuchon (2010) 0.04 ppb
26
Al
Robuchon (2010) 72 ppb
26
Al
Castillo-Rogez (2007)
A stripped sub-satellite
For the mass ratios which produce the largest
time-savings in de-spinning, q=0.060.021, the
sub-satellite is stripped by Saturn.
What happens to the sub-satellite?
In 90% of the cases it re-impacts Iapetus.
Will this increase Iapetus spin dramatically?
No. Likelihood is very small for impact to substantially
increase the spin rate for most q.
Topography
Giese et al. 2008
Iapetus has at least 7
basins between
300-800 km.
The sub-satellite re-
impact would impact
near escape velocity
~0.58 km/s.
Impactors with
q=0.0050.021
match the basin
sizes.
Building the ridge
After accreting out of a ring of debris the sub-
satellite can/will push the remaining debris to the
surface.
Tidal spreading times of a ring due to an external
perturber (sub-satellite) are very short ~10-1000s
years
The in-falling debris will have velocities nearly
tangential to the surface at only 300 m/s.
Will the morphology match all the details of the
observed ridge? (current work)
The Ridge?
Giese et al. 2008
Building a sandpile
Where the particles fall with ONLY tangential
velocity.
v
Lang
= 300 m/s
This sandpiles
properties will largely be
governed by its granular
properties
What will govern the
properties of this
sandpiles
properties????
Methods
We use pkdgrav to handle particle-particle
interactions
We model only a small patch of the surface with
periodic boundary conditions
Drop particles with 300 m/s tangential velocity
into the patch
v
x
y

y
y
v
tangential

Our experiment
Relative effects of ridge/sandpile construction
with tangential velocity
We really dont know the state of the in-falling
material, or the surface material, or the particle
size distribution, or the actual timescales.
Also, the numerical challenge of modeling a large
static clump/pile/ridge AND in-falling particles with
v=300m/s eliminates the possibility of modeling
the entire thing even in a patch.
Conclusions
A sub-satellite around Iapetus can decrease de-spinning
times by a factor of 10.
Sub-satellites which are stripped by Saturn re-impact
Iapetus 90% of the time.
The remnants of the debris disk, out of which the sub-
satellite forms, would be pushed to the surface, possibly
explaining the ridge of Iapetus.
1. Impact forms a ring of debris out of which a
satellite accretes
2. Ring gets pushed to the surface to form the
ridge
3. Tidal interactions between satellite and
Iapetus help to de-spin Iapetus, and push the
satellite to an orbit large enough to be
stripped by Saturn